人機交互之安全性:基於機器人相關標準概念和實現

Abstract

由於典型的工業機器人的體積、重量和高速以及人機交互的技術障礙，機器人被認爲是對人類操作者的安全具有較高風險的機器人。隨着機器人逐漸進入更多人居住的環境，他們將需要更好地獲取和解釋有關他們的環境的信息。以前通過將機器人與人類隔離而實現的人機交互的安全性將不再站得住腳。這意味着在HRC和pHRI領域，機器人的安全問題必須重新考慮，因此本論文致力於探討人機協作的安全問題。我們將從機器人安全標準的角度探討與機器人協作相關的安全策略。主要目標是提供有效和實用的安全策略，符合機器人安全標準。 因機器人安全標準屬於機械安全框架，我們將先討論機械安全標準。從風險管理的角度闡述了機械安全的核心概念和方法，還有與安全有關的控制系統的評價程序和方法。另外介紹了C類機器人的標準，包括ISO10218, ISO15066, ISO13482。描述並定義了工業環境下的四種HRC方法和要求，也為輔助機器人、個人護理機器人和服務機器人提供了安全要求。在瞭解了機器人相關安全標準的要求後，我們基於安全標準的核心概念提出了可行的安全策略，包含碰撞前及碰撞後安全策略。 關於碰撞前安全策略，我們根據ISO 15066標準提出了一種結合危險指數和虛擬阻抗控制的碰撞前安全策略。在此策略中，當操作員靠近機器人時，控制系統有能力使機器人減速。且當操作者持續靠近，使操作者與機器人之間的距離小於允許值時，則對機器人的軌跡進行修正以避免碰撞。該控制策略使機器人可以與人保持一定的距離，使操作者與機器人進行協作互動時能夠有效避免碰撞。 而碰撞後安全策略則是一種結合危險指數和強健式故障檢測與隔離(RFDI)的策略，並將其應用於主動-被動變剛度彈性執行器(APVSEA)。該APVSEA可以根據危險指數的變化主動改變關節剛度，且即使將關節剛度調整到最小值，也能爲用戶的安全提供額外的保護措施。

Due to the size, weight, and high speed of typical industrial robots and the technical hurdles in human–robot interaction, robots have been regarded as high risk in terms of the safety of human operators. As robots are gradually entering more human-populated environments, they will need to become better at acquiring and interpreting information about their environment. The safety of human–robot interactions previously achieved by isolating robots from humans will no longer be tenable. This means that in the fields of human robot collaboration (HRC) and physical human–robot interaction (pHRI), the issue of robot safety must be reconsidered. This dissertation is devoted to realize the safety of HRC. We will discuss the safety strategies relating to robot cooperation from the perspective of robot safety standards. And the main objective is to provide effective and practical safety strategies that are in compliance with robot safety standards. Since the robot safety standards are under the framework machinery safety, the safety standards for machinery will be discussed first. The core concept and methods of machinery safety and evaluation procedures for safety-related control systems are described on the basis of risk management. In addition, we describe the Type c standards of robots, including ISO10218, ISO15066, and ISO13482 and define the four HRC methods and requirements in an industrial environment. The safety requirements for assistive, personal care, and service robots were provided, too. After understanding the requirements of the safety standards related to robots, we put forward feasible safety strategies based on the core concepts of safety standards including pre-collision safety strategy and post-collision safety strategy. For the pre-collision safety strategy, we propose a pre-collision safety strategy in accordance with ISO 15066, which combines danger index and virtual impedance control. In this strategy, the control system has the ability to slow down a robot if an operator approaches it. If the operator continues approaching the robot and the distance between them is less than an allowable value, the original trajectory of the robot will be modified to avoid collisions. The control strategy decreases the speed of robots to keep a certain distance between robots and humans and allows the effective avoidance of collisions when human operators try to make physical contact with robots. Finally, the post-collision safety strategy combining a danger index and robust fault detection and isolation (RFDI) is presented and applied to an active–passive variable stiffness elastic actuator (APVSEA). This APVSEA can actively change joint stiffness with a change in the danger index and provide additional protection measures for the safety of users even if the joint stiffness is adjusted to the minimum value.